Development of Novel Monoclonal Antibodies Against CD109 Overexpressed in Human Pancreatic Cancer

www.oncotarget.com Oncotarget, 2018, Vol. 9, (No. 28), pp: 19994-20007

Research Paper Development of novel monoclonal antibodies against CD109 overexpressed in human pancreatic cancer

Gustavo A. Arias-Pinilla1, Angus G. Dalgleish2, Satvinder Mudan3, Izhar Bagwan4, Anthony J. Walker1 and Helmout Modjtahedi1 1School of Life Sciences, Pharmacy and Chemistry, Kingston University London, Kingston-upon-Thames, Surrey, UK 2Department of Cellular and Molecular Medicine, St George’s University of London, London, UK 3Department of Surgery of Hammersmith Campus, Imperial College, London, UK 4Department of Histopathology, Royal Surrey County Hospital, Guildford, UK Correspondence to: Helmout Modjtahedi, email: [email protected] Keywords: pancreatic cancer; monoclonal antibodies; CD109 antigen; tissue arrays; immunohistochemistry Received: December 10, 2017 Accepted: March 15, 2018 Published: April 13, 2018 Copyright: Arias-Pinilla et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License 3.0 (CC BY 3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

ABSTRACT

Pancreatic cancer is one of the most aggressive and lethal types of cancer, and more effective therapeutic agents are urgently needed. Overexpressed cell surface antigens are ideal targets for therapy with monoclonal antibody (mAb)-based drugs, but none have been approved for the treatment of pancreatic cancer. Here, we report development of two novel mouse mAbs, KU42.33C and KU43.13A, against the human pancreatic cancer cell line BxPC-3. Using ELISA, flow cytometry, competitive assay and immunoprecipitation followed by mass spectrometry, we discovered that these two mAbs target two distinct epitopes on the external domain of CD109 that are overexpressed by varying amounts in human pancreatic cancer cell lines. Treatment with these two naked antibodies alone did not affect tumour cell growth or migration in vitro. Of the two mAbs, only KU42.33C was useful in determining the expression of CD109 in tumour cells by Western blot and immunohistochemistry. Interestingly, immunohistochemistry of human pancreatic carcinoma tissue arrays with mAb KU42.33C showed that 94% of the 65 human pancreatic adenocarcinoma cases were CD109 positive, with no expression in normal pancreatic tissues. Our results suggest that these two novel mAbs are excellent tools for determining the expression level of CD109 in the tumour specimens and sera of patients with a wide range of cancers, in particular pancreatic cancer, and for investigating its diagnostic, prognostic and predictive value. Further research is warranted and should aim to unravel the therapeutic potential of the humanised forms or conjugated versions of such antibodies in patients whose tumours overexpress CD109 antigen.

INTRODUCTION Surgery is indicated in patients with localised disease, accounting only for 10-20% of cases, and is followed Pancreatic cancer is one of the most aggressive by adjuvant chemotherapy [2]. Treatment of metastatic and lethal types of cancer with a five-year survival rate disease is based on gemcitabine-containing regimens, of ~5% which has barely improved in the past four sometimes in combination with erlotinib or more decades [1]. In 2012, there were an estimated 337,872 recently with nab-paclitaxel although such combinations new cases and 330,391 deaths as a result of pancreatic only offer modest survival benefit in some patients in cancer worldwide [1]. Treatment of pancreatic cancer comparison with gemcitabine alone [3–5]. Treatment with requires a multidisciplinary approach that includes gemcitabine-free FOLFIRINOX (leucovorin, fluorouracil, surgery, chemotherapy, radiotherapy and palliative care. irinotecan, oxaliplatin) has improved the median overall www.oncotarget.com 19994 Oncotarget survival compared with gemcitabine alone (11.1 vs 6.8 KU42.33C and KU43.13A was found to be trypsin- months), but this combinational therapy is accompanied sensitive and there was no binding when screened by by increased side effects and thus is often an option for flow cytometry (data not shown). Therefore, an enzyme- younger patients with good performance status [6]. There free cell dissociation buffer was used for cell detachment is currently no reliable biomarker or screening method prior to flow cytometry analysis. Following optimisation, for the early detection of pancreatic cancer and novel and high level of expression of the target antigen was found more effective therapeutics are urgently needed. Moreover, with both mAbs KU42.33C and KU43.13A in some of without further advances in these fronts, pancreatic cancer the human pancreatic cancer cell lines including BxPC- is projected to bypass breast, prostate and colorectal 3 (MFI=86 and 90), Capan-2 (MFI=19 and 16), MIA cancers, and become the second leading cause of cancer PaCa-2 (MFI=56 and 57), PANC-1 (MFI=45 and 56), deaths after lung cancer in the USA by 2030, and the third CFPAC-1 (MFI=21 and 19) and FA-6 (MFI=39 and 36, leading cause of cancer death in European Union after respectively), compared to negative control (Figure 1, lung and colorectal cancer by 2025 [7, 8]. Supplementary Figure 2). In contrast, the expression of Monoclonal antibody (mAb) technology is the target antigen in other human pancreatic cancer cell an excellent tool for the identification of novel and lines namely Capan-1, AsPC-1, HPAF-II and Hs766T was overexpressed cell surface antigens in human malignancies found to be negative or lower with MFI values ranging and mAb-based products have untapped therapeutic from 2-12 (Figure 1, Supplementary Figure 2). We also and diagnostic potential in cancer. Indeed, the exquisite examined the level of expression of the target antigen in specificity of mAbs for their target antigens makes them other cancer cell lines including human head and neck an attractive approach for targeted cancer therapy [9–14]. squamous carcinoma HN5 and breast cancer cell lines To date, 32 mAbs have been approved for cancer treatment MDA-MB-468 and SKBR-3. High level of the target in the U.S. and/or European Union. However, none have antigen was present in both HN5 and the triple negative yet been approved for pancreatic cancer [15–18]. Our aim breast cancer cell line MDA-MB-468, but not in the HER- was to develop novel mAbs of diagnostic and therapeutic 2 overexpressing breast cancer cell line SKBR-3 (Figure potential against overexpressed cell surface antigens 1, Supplementary Figure 2). on human pancreatic cancer cells. Here, we report the Both novel mAbs were purified by affinity development of two novel mouse mAbs against distinct chromatography and were found to be of IgG1κ isotype epitopes on the extracellular domain of CD109, using the (Supplementary Figure 3). To determine whether the two human pancreatic cancer cell line BxPC-3 as the source of antibodies were directed against the same, overlapping or tumour immunogen. We showed that the CD109 antigen distinct epitopes on the target antigen, competitive ELISA is overexpressed in several human pancreatic cancer cell was performed. For this, saturation of the antigen binding lines and pancreatic cancer tissue array, with no expression site was initially performed with one the mAbs prior to in normal pancreatic tissues. The characterisation of these addition of the second antibody to explore if absorbance antibodies and their potential for use as diagnostic and value increased further. The absorbance curve for mAb therapeutic agents are discussed. KU42.33C plateaued at concentrations higher than 50 μg/ ml. Thereafter, only additional doses of mAb KU43.13A RESULTS increased the absorbance (Figure 2). These results support that mAbs KU42.33C and KU43.13A recognise non- Development of two novel mouse mAbs overlapping epitopes on the extracellular domain on the KU42.33C and KU43.13A target antigen.

Several hybridoma were generated by the fusion of The two novel mAbs KU42.33C and KU43.13A lymphocytes, isolated from the spleens of mice immunised are against CD109 on human pancreatic cancer with BxPC-3 cells, and the myeloma cell line SP2. The cells hybridomas’ supernatants were screened against a panel of human pancreatic cancer cell lines established from Both mAbs, KU42.33C and KU43.13A, patients at different stages of their disease including immunoprecipitated a ~170 KDa protein identified as primary tumours (BxPC-3, Capan-2, MIA PaCa-2, PANC- CD109 by mass spectrometry (Figure 3A and Table 1), liver metastatic tumours (CFPAC-1, Capan-1), ascites 1). In addition, mAb KU42.33C immunoprecipitated (HPAF-II, AsPC-1) and lymph node metastasis (Hs 766T). faint bands of ~150 KDa and ~80 KDa (Figure 3A). The results of ELISA screening showed that the antibodies However, in Western blot, only KU42.33C detected the secreted by two hybridomas, KU42.33C and KU43.13A, CD109 bands of 170 KDa and 150 KDa, suggesting that were against an antigen with high level of expression in KU43.13A is directed against a conformational epitope some of the human pancreatic cell lines (e.g. BxPC-3, on CD109 antigen (Figures 3B). Furthermore, to confirm PANC-1, FA-6 and MIA PaCa-2, Supplementary Figure that these two antibodies were directed against CD109, 1). However, the target antigen recognised by mAbs BxPC-3 tumour lysates were immunoprecipitated with www.oncotarget.com 19995 Oncotarget the two novel antibodies, transferred by Western blot and 4A & 4B). The anti-EGFR mAb HM43.16B was used as then probed with a commercial anti-CD109 antibody. As a control and treatment with this antibody resulted in the shown in Figure 3C, the commercial antibody detected EGFR downregulation and internalisation (Figure 4A). the CD109 antigen immunoprecipitated by both of the novel mAbs. These results suggest that mAbs KU42.33C The application of mAb KU42.33C in and KU43.13A are directed towards a sequential and a immunohistochemical detection of CD109 in conformational epitope on CD109 antigen, respectively. pancreatic cancer Effect of novel mAbs KU42.33C and KU43.13A To explore the diagnostic potential of the novel on tumour growth and migration and cellular mAbs, immunohistochemical staining of CD109 positive location of CD109 BxPC-3 tumour cell pellets was investigated. Our results showed that only mAb KU42.33C, which was found to be Next, we examined the effect of treatment with the against a sequential epitope by Western blot analysis as two novel antibodies on growth and migration, in vitro, of described above, detected the antigen in formalin-fixed, a panel of human pancreatic cancer cell lines. At maximum paraffin-embedded tissue sections (Figure 5A and 5B). concentration of 300 nM, mAbs KU42.33C and KU43.13A Further optimisation, including antigen retrieval with Tris- did not affect the growth of the human pancreatic cancer EDTA pH 9.0 buffer and signal amplification, resulted in cell lines nor inhibited the migration of BxPC-3, AsPC-1 stronger immunostaining of CD109 (data not shown). and CFPAC-1 cells (data not shown). We also examined Next, we examined the relative expression of CD109 the cellular location of the target antigens recognised by antigen by mAb KU42.33C in human pancreatic cancer the mAbs. As shown in Figure 4A, the CD109 antigen tissue arrays containing 70 specimens from patients with recognised by these novel mAbs is located on the surface pancreatic adenocarcinoma. Of these, five samples were of human pancreatic cancer cells. In addition, there was either necrotic or contained no tumour and therefore were no evidence of CD109 downregulation and internalisation excluded from the study. Of 65 cases, 94% were CD109 following antibody treatment for 30 min at 37ºC (Figures positive, with intensity ranging from 1+ (n=55, 85%) to 2+

Figure 1: Expression level of the antigens recognised by novel mAbs KU42. 33C and KU43.13A on human pancreatic cancer and other cancer cell lines, determined by flow cytometry. Data is presented as Mean Fluorescence Intensity (MFI) ± SD. www.oncotarget.com 19996 Oncotarget (n=6, 9%). Staining predominated in the membrane and mAb technology is an excellent tool for the discovery of the cytoplasm of cancer cells and did not correlate with novel overexpressed cell surface tumour antigens and the the disease grade (Figure 5C-5G; Supplementary Table study of their functions. Here, we report the generation 1). In contrast, normal pancreatic tissue was negative of two mAbs, KU42.33C and KU43.13A, against two or showed weak diffuse staining possibly attributable to distinct epitopes on the external domain of CD109, using technical artefact (Figure 5H). There was no staining in the human pancreatic cancer cell line BxPC-3 as the the absence of primary antibody-control (data not shown). source of tumour immunogen. We have shown that the Interestingly, with the exceptions of weak staining (1+ antigen recognised by these antibodies is overexpressed intensity) of normal tissues from ovary and small intestine, on the cell surface of BxPC-3 and other human pancreatic all the remaining 31 normal tissues from the following cancer cell lines by varying amounts (Figures 1 and 4A). organs were CD109 negative: cerebellum, cerebral cortex, However, treatment with these two antibodies did not pituitary gland, spinal cord, eye, adrenal gland, thyroid, affect the proliferation or migration in vitro of CD109 parathyroid adenoma, thymus, tonsil, bone marrow, kidney, overexpressing cancer cell lines (data not shown), bladder, prostate, testis, uterine cervix, endometrium, nor resulted in the down-regulation of CD109 antigen fallopian tube, oesophagus, stomach, colon, rectum, liver, (Figure 4). Interestingly, of the two antibodies, only mAb pancreas, spleen, lung, heart, breast, placenta, striated KU42.33C recognised the CD109 antigen by Western blot muscle and skin (Supplementary Table 2). (Figure 3B), which suggests that both mAbs are directed against different epitopes of the same antigen (Figure 2). DISCUSSION Originally identified by mAbs raised against the primitive CD34+ acute myeloid leukaemia cell line Pancreatic cancer is projected to become the KG1a, CD109 is a transforming growth factor (TGF)-β second leading cause of cancer deaths in the USA by co-receptor that binds TGF-β1, regulates TGF-β receptor 2030 [7]. There is an urgent need to discover novel endocytosis and degradation, and suppresses TGF-β/ biomarkers of diagnostic, prognostic and predictive Smad signalling [19–23]. TGF-β has been found to values and therapeutic targets in patients with pancreatic play a role in growth, differentiation and migration; cancer. As mAb-based products are highly specific for dysregulation of TGF-β signalling is associated with their target antigens, they have been routinely used as tissue fibrosis and cancer [21]. CD109 is a monomeric diagnostic and therapeutic agents in many pathological 170 KDa glycosylphosphatidylinositol (GPI)-anchored conditions including cancer [9–12, 17–18]. Moreover, cell surface protein, a member of the α2-macroglobulin/

Figure 2: MAbs KU42.33C and KU43.13A are directed against distinct epitopes on BxPC-3 human pancreatic cancer cells. BxPC-3 cells were grown in 96-well plates to near confluency. MAb KU42.33C was added in doubling dilutions at highest concentration of 25 μg/ml for 1 hour on ice. After washing unbound antibodies, increasing concentrations of mAb KU42.33C or mAb KU43.13A were added and the total level of bound antibodies were determined as described in Materials and Method section. Data is presented as mean absorbance at 450 nm ± SD. www.oncotarget.com 19997 Oncotarget Table 1: Identification of proteins recognised by novel mAbs KU42.33C and KU43.13A by mass spectrometry Protein Hits Band No. mAb Matches/Sequences Q6YHK3 CD109 antigen OS=Homo sapiens GN=CD109 PE=1 SV=2 1 KU42.33C Mass: 162500 Score: 381 Matches: 6(6) Sequences: 6(6) Start-End Score Peptide 375–385 45 K.LSDSWQPR.S 649–663 48 K.FLIDTHNR.L 859–872 64 R.ADGNQLTLEER.R 879–894 99 K.SYSQSILLDLTDNR.L 1131– 53 K.TLSFSFPPNTVTGSER.V 1138 1220– 73 K.DYIDGVYDNAEYAER.F 1227 Q6YHK3 CD109 antigen OS=Homo sapiens GN=CD109 PE=1 SV=2 2 KU43.13A Mass: 162500 Score: 158 Matches: 3(3) Sequences: 3(3) Start-End Score Peptide 375–385 62 R.ADGNQLTLEER.R 1131– 42 K.LSDSWQPR.S 1138 1220– 54 K.FLIDTHNR.L 1227

C3, C4, C5 family of thioester-containing proteins [19, These findings are consistent with the results obtained 21, 24]. CD109 is found on the cell surface of activated with our two anti-CD109 antibodies using ELISA and platelets and T-cells, endothelial cells and a subpopulation flow cytometry (Figure 1, Supplementary Figures 1 & of CD34+ haematopoietic and progenitor cells, but is 2). They also examined CD109 expression in 18 tissue not expressed in most normal human tissues except for sections from pancreatic ductal adenocarcinoma (PDAC) myoepithelial cells of mammary, lacrimal and salivary and 11 normal pancreatic tissue samples and found glands, and basal cells of bronchial and prostate epithelia positive IHC CD109 staining of variable intensity in [19, 20, 23, 25-27]. In addition to the 170 KDa major pancreatic carcinoma cells and completely negative or CD109 protein band, the identification of a 150 KDa rare cases of focal and weak immunoreactivity in normal band has been reported and found due to the proteolytic/ pancreatic tissue. A substantial difference in CD109 autolytic cleavage of the 170 KDa counterpart [19, 25]. expression in pancreatic adenocarcinoma compared In agreement, we found that of our two novel antibodies, to normal pancreatic tissue was also observed and no mAb KU42.33C, which targets a sequential epitope on the staining was seen in other pancreatic tissue components external domain of CD109, detected both the 170 KDa such as blood vessels, pancreatic acini, stromal fibrous and 150 KDa protein bands by immunoprecipitation and tissue, adipose tissue and inflammatory cells [26]. In our Western blot (Figure 3A and 3B). study, we examined pancreatic cancer tissue arrays of There is currently no comprehensive study samples from 65 patients with different tumour grades on the expression pattern of CD109 determined by and 94% of the cases were CD109-positive; we found immunohistochemistry and its prognostic significance and no correlation between the staining intensity and tumour predictive value for response to therapy in patients with grade. The staining predominated in the cytoplasm of pancreatic cancer. In only one study, Haun and colleagues cancer cells although there some cases of coexisting examined the expression of CD109 in a panel of eight membrane staining (Figures 5C-5G; Supplementary Table human pancreatic cancer cell lines by Western blot. They 1). Interestingly, the antibody did not stain normal human found high levels of expression in BxPC-3, MIA PaCa- pancreatic tissue (Figure 5H). No staining was seen in a 31 2, and PANC-1 cells, with no/low expression in A818- out of 33 normal organs in tissue arrays (Supplementary 4, AsPC-1, Capan-1, CFPAC-1 and Suit-2 cells [26]. Table 2). www.oncotarget.com 19998 Oncotarget Interestingly, CD109 has been found overexpressed of lower-grade glioma and therefore a potential molecular in a number of other cancer types including carcinoma target for therapy [36, 38]. of the uterine cervix [28], lung squamous cell carcinoma There is currently conflicting data on the [20], cancer of the oral cavity [29], breast cancer [30– prognostic significance CD109 expression in such 31], malignant melanoma [32], myxofibrosarcoma [23], cancers. For example, CD109 expression determined by esophageal squamous cell carcinoma [22], squamous immunohistochemistry in tumour specimens from patients cell/adenosquamous carcinomas of the gallbladder with myxofibrosarcoma and hepatocellular carcinoma was [33], hepatocellular carcinoma [34], nasopharyngeal associated with a poorer prognosis in such patients [23, carcinoma [35] and penile squamous cell carcinoma 34]. The expression of CD109 has also been reported in [36]. For example, high levels of CD109 expression were cancer stem-like cells/cancer-initiating cells (CSCs/CICs) observed by immunohistochemistry in squamous cell in the epithelioid sarcoma cell line ESX, and was found carcinomas and premalignant lesions of the oral cavity to be associated with poor prognosis in patients with soft but not in normal tissue, suggesting that CD109 is useful tissue sarcoma [39]. Additionally, a significant association in predicting transformation of premalignant lesions into was observed between high CD109 expression and low cancer [29]. CD109 expression was also found to be more 1-year survival in patients with diffuse large B-cell frequent in lung squamous cell carcinomas compared with lymphoma [40]. In contrast, CD109 expression was other types of lung carcinoma including adenocarcinomas, associated with better prognosis in patients with urothelial large cell carcinomas and small cell carcinomas; however, carcinomas [27]. CD109 expression was also found to be no association between CD109 expression and the clinical greater in triple-negative breast cancer (TNBC) compared stage of the disease was found [20]. Similarly, Dong and to non-TNBC and to be associated with higher rate of co-workers showed that CD109 is expressed in bladder distant metastasis and worse postoperative disease-specific squamous cell carcinomas and adenosquamous carcinomas survival compared to those with low or no expression but not in adenocarcinomas or normal gallbladder tissue [31]. In agreement, we also found CD109 overexpression [33]. A positive correlation has been reported between in the triple negative EGFR overexpressing breast cancer CD109 expression and tumour grade in patients with cell line MDA-MB-468, but not in the ER-, PR-, HER2 vulvar squamous cell carcinoma [37]. More recently, overexpressing cell line SKBR-3 (Figure 1; Supplementary CD109 was found highly expressed in penile squamous Figure 2). We also found overexpression of CD109 in cell carcinoma and to be a key regulator of the progression other EGFR overexpressing head and neck squamous

Figure 3: Immunoprecipitation and immunodetection by Western blot of CD109 antigen with novel mAbs KU42.33C and KU43.13A. (A) Immunoprecipitation was performed with novel mAbs KU42.33C and KU43.13A using sheep anti-mouse dynabeads. A protein band of ~170 KDa was immunoprecipitated with both novel mAbs. The ~50/25 KDa bands represent heavy and light chains of the anti-mouse antibody, respectively. (B) CD109 antigen was immunoprecipitated and probed with novel mAbs KU42.33C and KU43.13A. CD109 was immunodetected by mAb KU42.33C but not by mAb KU43.13A. (C) CD109 antigen was immunoprecipitated with mAbs KU42.33C and KU43.13A and subsequently immunodetected with commercial anti-CD109 mAb. MWM: molecular weight marker. www.oncotarget.com 19999 Oncotarget Figure 4: Internalisation studies of novel mAbs KU42.33C and KU43.13A in BxPC-3 human pancreatic cancer cells determined by (A) immunofluorescence, and (B) ELISA. BxPC-3 cancer cells were grown to near confluency and incubated with purified antibodies (50 μg/ ml) or control (PBS/1% BSA) at 4ºC for 1 h and subsequently at 37ºC for extra 30 min to allow internalisation. Cells were then fixed, permeabilised and incubated with FITC-conjugated or HRP-linked anti-mouse secondary antibodies for immunofluorescence staining (200x magnifications) and ELISA respectively. The anti-EGFR mAb HM43.16B. was used as a control (arrows: internalised antibody). ELISA results (B) are presented as mean absorbance ± SD. www.oncotarget.com 20000 Oncotarget cell carcinoma cell line HN5 (Figure 1; Supplementary therapeutic benefit in the neoadjuvant or adjuvant setting Figure 2). More recently, CD109 was reported to act as [41]. While the soluble recombinant form of CD109 a novel pro-metastatic factor in a lung adenocarcinoma was reported to be capable of binding to TGF-β and mouse model, and high levels of CD109 resulted in the antagonising TGF-β signalling and cellular responses in activation of Jak-Stat3 signalling pathways in cancer cells experimental system [42, 43], it remains unclear whether suggesting that the direct targeting of CD109 could be of patients with pancreatic cancer or any other cancer type

Figure 5: Examples of immunohistochemical staining of formalin-fixed, paraffin-embedded BxPC-3 cancer cell pellets (A-B) and pancreatic cancer tissue arrays (C-H) using novel mAbs KU43.13A (5 μg/ml; A) and KU42.33C (5 μg/ml; B-H). Negative and strong staining of BxPC-3 cell pellets with mAb KU43.13A (A) and mAb KU42.33C (B) respectively (magnification 200x). Representative images of staining of pancreatic cancer specimens in tissue arrays: (C) 1+ cytoplasmic staining, magnification 200x; (D) 1+ cytoplasmic/ membranous staining, magnification 200x; (E) 2+ cytoplasmic/membranous staining, magnification 200x; (F) 2+ cytoplasmic/membranous staining, magnification 400x; (G) 2+ membranous staining with perineural invasion, magnification 400x; (H) Negative staining in normal pancreas, magnification 200x. www.oncotarget.com 20001 Oncotarget shed CD109 antigen into their sera and any other body The EGFR overexpressing head and neck squamous cell fluids and if so, whether CD109 may confer diagnostic, carcinoma (HN5) and the breast carcinoma cell lines prognostic and predictive values, and therefore warrants MDA-MB-468 and the HER-2 overexpressing SKBR-3 further investigation. Finally, while TMAs are commonly were also used as described previously [48]. MIA PaCa-2, employed in medical research, the heterogeneous nature of PANC-1, Capan-1, MDA-MB468 and SKBR3 cells were both tumour and stroma in patients with pancreatic cancer authenticated by LGC Standards (Teddington, UK). support the need for further investigations on the relative All cancer cell lines were cultured routinely at expression and prognostic significance of CD109 in a 37ºC in a humidified atmosphere with 5% CO2. BxPC- larger group of patients using the whole tumour sections 3, AsPC-1, Capan-1 and FA-6 were cultured in RPMI- or several TMA cores from the same tumours [44, 45] 1640 medium, MIA PaCa-2, PANC-1 and Hs 766T In summary, our results suggest that high expression in Dulbecco’s Modified Eagle’s Medium, HPAF-II in of CD109 is common in pancreatic cancer. We believe Eagle's Minimum Essential Medium, CFPAC-1 in Iscove's that our two novel anti-CD109 mAbs, which are directed Modified Dulbecco's Medium and Capan-2 in McCoy's against two distinct epitopes on the extracellular domain 5a Medium Modified. Cells were supplemented with 10% of CD109, are ideal tools for conducting detailed studies Foetal Bovine Serum (FBS) and antibiotics penicillin (50 of the biological significance and diagnostic, prognostic units/ml), streptomycin (0.05 mg/ml) and neomycin (0.1 and predictive values of CD109 in human cancers. mg/ml). RPMI-1640 medium, Eagle's Minimum Essential To our knowledge, there is currently no study on the Medium and Iscove's Modified Dulbecco's Medium were importance of CD109 as target for therapy with mAbs in supplemented with 2 mM, 4 mM and 8 mM L-Glutamine pancreatic cancer or any other type of cancer. Moreover, respectively. MDA-MB-468, SKBR-3, HN5 and SP2 cells at present, the great majority of FDA approved therapeutic were grown in Dulbecco’s Modified Eagle’s Medium antibodies against tumour associated antigens (e.g. anti- supplemented with 10% FBS and antibiotics. All culture CD20, anti-CD30, anti-EGFR, anti-HER-2 mAbs) are media and additives were purchased from Sigma Aldrich, the humanised/chimeric IgG1 version of the mouse UK. mAbs due to their superior ADCC (antibody-dependent cell-mediated cytotoxicity) and/or CDC (complement- Generation of novel monoclonal antibodies dependent cytotoxicity) functions, or conjugated to toxins and radioactive substances to deliver lethal doses of such The mice immunisation was performed at St agents to the tumour. A major reason being that, with the George’s University of London, following the ethical exception of growth factor receptor blocking antibodies approval and under Home Office animal license. Female (e.g. anti-EGFR mAbs), treatment with the majority BALB/c mice (aged 5-6 weeks) were immunised of naked mAbs do not inhibit the growth in vitro and repeatedly with subcutaneous flank injections plus migration of tumours overexpressing the target antigens [9, an intraperitoneal (i.p.) injection of BxPC-3 human 17–18, 46]. As treatment with our two anti-CD109 mAbs pancreatic cancer cells (3 sites; 100 μl per site) with a total did not affect the proliferation or migration of CD109 number of 10 million tumour cells per immunisation per overexpressing cancer cell lines, further investigations mouse. Immunisation was repeated 2 times every 2 weeks are warranted to determine the therapeutic potential of and the final injection was administered 3-4 days before the humanised IgG1 or the conjugated versions of such collection of lymphocytes from the spleen of immunised antibodies, for use in targeted therapy of tumours with mice. Spleens were removed from the immunised animals CD109 overexpression. by blunt dissection, disaggregated, and the cells washed, centrifuged and resuspended in a freezing down solution MATERIALS AND METHODS for storage in liquid nitrogen until the date of fusion. B-lymphocytes derived from the spleen of immunised mice were fused with SP2 myeloma cells, Cancer cell lines and cell culture facilitated by 50% polyethylene glycol (PEG; Sigma Aldrich). Cells were cultured in HAT medium (Sigma A panel of ten human pancreatic cancer cell lines Aldrich) supplemented with 20% FBS, 10% Hybridoma was used in this study: BxPC-3, Capan-2, MIA PaCa-2, Cloning Supplement (Santa Cruz Biotechnology, UK) and PANC-1, AsPC-1, HPAF-II, CFPAC-1, Capan-1, Hs766T antibiotics. Newly formed hybridomas were screened by and FA-6 [47]. BxPC-3, PANC-1, MIA PaCa-2 and SP2 ELISA, cloned twice by limiting dilution technique, grown myeloma cells were purchased from European Collection in roller bottles and the hybridoma supernatants were of Cell Cultures (ECACC, UK). AsPC-1, Capan-2, harvested. Isotyping of novel mAbs was determined using HPAF-II, Hs766T and CFPAC-1 were purchased from a mouse mAb isotyping kit (AbD Serotec, UK) according American Type Culture Collection (ATCC, UK). Capan-1 to the manufacturer’s protocol and the antibodies were was kindly provided by Dr Charlotte Edling (Barts and purified by affinity chromatography (below). The London School of Medicine and Dentistry) [47]. www.oncotarget.com 20002 Oncotarget Antibody screening by ELISA and flow Competitive binding assay cytometry analysis BxPC-3 cancer cells were seeded in 96-well plates Antibodies secreted by novel hybridomas were and incubated until they became near confluent. Cells screened by ELISA. Briefly, cancer cells were seeded were washed once with serum-free DMEM and increasing in 96-well plates and incubated until near confluent. amounts of mAb KU42.33C were added to different Cells were washed once and then incubated with novel wells in duplicates until near saturation point. Thereafter, hybridoma supernatants or purified mAbs (25 μg/ml) increasing concentrations of either mAb KU42.33C or on ice for 1 h, followed by incubation with 50 μl HRP- KU43.13A were added in duplicates and incubated on conjugated rabbit anti-mouse secondary antibody (1:1000, ice for 1 hour. Following washing, 50 μl HRP-conjugated STAR13B, AbD Serotec) on ice for 45 min. Cells were rabbit anti-mouse secondary antibody (1:1000, STAR13B, washed and then 50 μl TMB (3,3',5,5'-Tetramethylbenzi AbD Serotec) was added and incubated on ice for 45 min. dine; Sigma Aldrich) added to each well and left for ~15 Cells were washed again three times and 50 μl TMB added min at room temperature; the TMB reaction was stopped to each well and left for ~15 min at room temperature; the TMB reaction was stopped by adding 50 μl of 0.5 M by adding 50 μl of 0.5 M sulphuric acid (H2SO4) to each well. Positive and negative controls were included. The sulphuric acid (H2SO4) to each sample and its absorbance absorbance of each sample was measured at 450 nm using measured at 450 nm. a Biotek plate reader (Biotek, UK). The cell surface expression of target antigens Internalisation studies recognised by novel mAbs was determined using flow cytometry as described previously [49]. Briefly, cells were Immunofluorescence staining of tumour cells and dissociated from the tissue culture flasks using enzyme- ELISA were used to determine whether treatment with free cell dissociation buffer (Fisher Scientific, UK). novel antibodies resulted in down-regulation of the target Approximately 1x106 tumour cells were incubated with antigen. Briefly, BxPC-3 cancer cells were grown to near novel mouse mAbs KU42.33C or KU43.13A (10 μg/ml) confluency in RPMI/10% FBS in Lab-Tek 8-well chamber or control (i.e. PBS) by rotation for 1 h at 4ºC, followed slides (VWR, UK) or 96-well plates, respectively. Cells by incubation with FITC-conjugated goat anti-mouse IgG were incubated with mAbs KU42.33C or KU43.13A (50 secondary antibody (1:200; STAR9B, AbD Serotec) for 45 μg/ml), or negative control (i.e. PBS/1%BSA alone) for 1 min at 4ºC. A minimum of 10,000 events were recorded h at 4ºC to allow antibody binding, followed by incubation by excitation with an argon laser at 488 nm and analysed at 37ºC for 30 min to allow antibody internalisation. The using the FL-1 detector (FITC detector; 525 nm) of a BD anti-EGFR mAb HM43.16B was used as another control FACScalibur flow cytometer using CellQuest Pro software in this study [50]. A control slide was maintained at 4ºC. (Becton-Dickinson Ltd, UK). Cells were fixed with 4% formaldehyde for 10 min, the cell membrane permeabilised with 0.5% Triton-X Purification of novel monoclonal antibodies 100 for 15 min and non-specific binding blocked with PBS/1%BSA for 1 h at 4ºC. Cells in immunofluorescence The purification of mAbs was performed by slides were then incubated with Alexa Fluor 488 secondary affinity chromatography. Briefly, novel mouse mAbs antibody (1:200; Fisher Scientific) for 1 h at 4ºC, mounted were purified by salt fractionation (solid ammonium in Vectashield with DAPI (Vector laboratories, UK) and examined using Nikon eclipse i80 microscope and Nikon sulphate ([NH4]2SO4; 45% of saturation - 270 g/L; Fisher Scientific) followed by affinity chromatography using a 5 NIS-Elements software as described previously [51]. Cells ml HiTrap Protein G HP column in an ÄKTAprime plus in ELISA plates were incubated with HRP-conjugated chromatography system (both from GE Healthcare, UK). rabbit anti-mouse secondary antibody and the absorbance The antibody sample was loaded into the HiTrap Protein G of each sample measured at 450 nm. HP column pre-equilibrated with 25 ml binding buffer (0.1 M phosphate buffer, 0.15 M NaCl, pH 7.4) at a flow rate of Cell growth studies 1 ml/min. Unbound proteins were removed by washing the column with 50 ml binding buffer and bound antibodies The effect of novel mAbs on the growth of were eluted with 25 ml elution buffer (0.1 M glycine, human cancer cell lines was investigated using the pH 2.5). Selected fractions were pooled together and Sulforhodamine B (SRB) colorimetric assay [47, 49]. 3 subjected to buffer exchange with PBS using the HiPrep Approximately 5x10 tumour cells per well were seeded 26/10 Desalting column (GE Healthcare Life Sciences). in 96-well plates containing medium supplemented with Following buffer exchange, the purified antibodies were 2% FBS in the presence of mAbs (300 nM) or control filtered through a 0.2 μM syringe filter (Merck Millipore, medium. AsPC-1 and HPAF-II cancer cells were grown UK), aliquoted and stored at -20ºC for further studies. in 5% FBS medium as they did not grow well in 2% FBS medium. A (no treatment) control plate was included under

www.oncotarget.com 20003 Oncotarget the same conditions (4 h incubation at 37ºC) to determine [10x] and 65% distilled water; Invitrogen, UK), heated to the initial number of cells before treatment. Cells were 95ºC for 5 min and analysed by SDS-PAGE. cultured until the controls became almost confluent (4-10 Identification of isolated protein was performed days). Tumour cells were fixed with 10% trichloroacetic by mass spectrometry at the University of York. Desired acid (TCA; Fisher Scientific), stained with 0.4% SRB bands were excised and in-gel digested with trypsin. (Sigma Aldrich) in 1% acetic acid and solubilised with Positive-ion MALDI mass spectra were obtained using a Tris-base (10 mM, pH 10; Fisher Scientific) before plate Bruker ultraflex III in reflectron mode, equipped with a reading. Gen5 software was used to determine the IC50 Nd:YAG smart beam laser. MS spectra were acquired over through non-linear least squares curve fitting [47, 49]. a mass range of m/z 800-5000. Final mass spectra were externally calibrated against an adjacent spot containing Migration assay 6 peptides (des-Arg1-Bradykinin, 904.681; Angiotensin I, 1296.685; Glu1-Fibrinopeptide B, 1750.677; ACTH The effect of novel mAbs on the migration of (1-17 clip), 2093.086; ACTH (18-39 clip), 2465.198; human pancreatic cancer cells BxPC-3, AsPC-1 and ACTH (7-38 clip), 3657.929). Monoisotopic masses CFPAC-1 was investigated using the IncuCyte ZOOM® were obtained using a SNAP averaging algorithm (C Live-Cell Imaging instrument (Essen Bioscience, UK) 4.9384, N 1.3577, O 1.4773, S 0.0417, H 7.7583) and a as described previously [52]. Approximately 1x103 S/N threshold of 2. For each spot, the ten strongest peaks cancer cells per well were seeded in duplicates in 0.5% of interest, with a S/N greater than 30, were selected for FBS medium in an IncuCyte™ ClearView 96-well Cell MS/MS fragmentation. Fragmentation was performed in Migration Plate (Essen Bioscience, UK) along with 300 LIFT mode without the introduction of a collision gas. The nM mAbs or control (i.e. medium alone). The cells were default calibration was used for MS/MS spectra, which allowed to settle for 15 min at room temperature before were baseline-subtracted and smoothed (Savitsky-Golay, transfer to an incubator at 37ºC for 30 min to pre-incubate width 0.15 m/z, cycles 4); monoisotopic peak detection the cells in the presence of treatment. Then, 200 μl of used a SNAP averaging algorithm (C 4.9384, N 1.3577, chemoattractant (i.e. 10% FBS medium) or control (i.e. O 1.4773, S 0.0417, H 7.7583) with a minimum S/N of 0.5% FBS medium) were added to the appropriate wells 6. Bruker flexAnalysis software (version 3.3) was used to of the reservoir plate and the insert plate placed into the perform the spectral processing and peak list generation pre-filled reservoir plate. The plate was then transferred for both the MS and MS/MS spectra. Tandem mass to the IncuCyte ZOOM® Live-Cell Imaging Instrument spectral data were submitted to database searching using (Essen Bioscience) and allowed to warm to 37ºC for 15 a locally-running copy of the Mascot program (Matrix min before any condensation accumulated on the plate Science Ltd., version 2.1), through the Bruker BioTools lid or bottom was wiped away. The plate was imaged at interface (version 3.2). Search criteria included: Enzyme, 10x objective using the Chemotaxis Scan Type - Phase Trypsin; Fixed modifications, Carbamidomethyl (C); channel. The IncuCyte™ Chemotaxis Cell Migration Variable modifications, Oxidation (M); Peptide tolerance, Software Module (Essen Bioscience) was used for data 100 ppm; MS/MS tolerance, 0.5 Da; Instrument, MALDI- analysis. Whole-well images of cells on both the bottom TOF-TOF. UniProt_human_SP database was used for and the top of the plate membrane were captured every protein identification. 2 h over 48 h and all images were processed using automatic algorithms to quantify cell area on each side of Western blotting the membrane. The ability of the novel antibodies to recognise the Immunoprecipitation and mass spectrometry target antigen in Western blot was performed. Briefly, protein immunoprecipitated with novel mAbs KU42.33C To identify the target antigens recognised by the and KU43.13A (5 μg) was analysed by SDS-PAGE under antibodies, immunoprecipitation and mass spectrometry reducing conditions, prior to Western blotting. The transfer were performed. Briefly, novel mAbs (5 μg) were of proteins from 4-12% Bis-Tris-gels to Immobilon- incubated with 1 ml BxPC-3 tumour cell lysates (lysis FL PVDF membranes (Merck Millipore, UK) was buffer containing 50 mM Tris-HCl pH 7.2, 150 mM NaCl, performed using the XCell II™ Mini-Cell Blot Module

2 mM MgCl2, 2 mM CaCl2, 0.1% NaN3, 100 mM DTT, kit (Invitrogen, UK) at a constant voltage of 30 V on ice 1% Triton X-100, 50 mM N-ethylmaleimide) overnight at for 2 h. PVDF membranes were probed with novel mAbs 4ºC by gentle rotation (14 rpm), and then incubated with (30 μg/ml) or commercial anti-CD109 mAb (1:100; sc- 50 μl pre-washed Dynabeads sheep anti-mouse IgG for 365780; Santa Cruz Biotechnology, UK) overnight at 4ºC 1 h at 4ºC. The immunocomplexes were captured on a and subsequently incubated with secondary goat anti- DynaMag™-2 for 2 min, the supernatants aspirated and mouse antibody (1:10,000; LI-COR Biosciences, UK) for the samples washed 3 times with PBS. The complexes 1 h at room temperature. For visualisation, the blots were were then eluted by mixing beads with LDS sample buffer analysed using the Oddysey® CLx instrument (LI-COR (25% NuPAGE LDS buffer [4x], 10% reducing agent Biosciences). www.oncotarget.com 20004 Oncotarget Immunohistochemical examination of tumours AW co-supervised the PhD project. HM, AGD, AW, SM, and IB helped with technical training and data analysis. BxPC-3 cell pellets were used to determine whether GAP and HM wrote the paper and AGD, AW, SM, and IB novel antibodies can be used for detection of the target edited and approved the final version of the manuscript. antigen in formalin-fixed paraffin-embedded tissues and for further optimisation of the primary antibodies. ACKNOWLEDGMENTS Then, the expression level of the target antigens was determined in human pancreatic cancer tissue arrays (35 This work was supported by Kingston University cases in early stage, 35 cases in advance stage and 10 London towards a PhD project and then The Ralph Bates cases normal tissue; Cat. No. PA804a, Biomax US) and Pancreatic Cancer Research Fund’s Postdoctoral grant. normal organ tissue arrays (33 types of normal tissue; Cat. We thank Dr. Wai Liu (St George’s University London) No. MNO661, Biomax US) by the mAb. Briefly, tissue and Dr. Daniel Fowler (St George’s University of London, sections were deparaffinised and rehydrated through a Funded by Institute of Cancer Vaccine and Medicine) for series of alcohols followed by antigen retrieval with Tris- conducting the immunisation procedure. EDTA buffer (10 mM TrisBase, 1 mM EDTA, 0.05% Tween 20, pH 9.0). Then, the mouse-specific HRP/DAB Ethics statement (ABC) Detection IHC Kit (Cat. No. ab64259, Abcam, UK) was used according to the manufacturer instructions. All applicable international, national, and/or Briefly, tissue sections were incubated with hydrogen institutional guidelines for the care and use of animals peroxide block for 10 min, followed by protein block for were followed. All procedures performed in studies 5 min and subsequent incubation with novel mAbs (5 involving animals were in accordance with the ethical μg/ml) overnight at 4 °C. Following washing, the tissue standards of St. George’s University of London at which sections were incubated with biotinylated goat anti-mouse the studies were conducted. secondary antibody for 10 min, followed by streptavidin peroxidase for another 10 min. After washing, the CONFLICTS OF INTEREST colorimetric reaction was developed by incubation with DAB substrate plus chromogen for 5 min. The slides were The authors report no conflicts of interest. counterstained with Harris Haematoxylin (VWR, UK) for 2 min and rinsed with water. The tissue sections were then dehydrated through a series of alcohols and the slides FUNDING mounted in DPX mounting medium (VWR) and examined for the expression level and intensity of staining (i.e. 1+ This work was supported by Kingston University weak, 2+ moderate, 3+ strong). London, United Kingdom and The Ralph Bates Pancreatic Cancer Research Fund. Statistical analysis REFERENCES Statistical analysis was performed using the Statistical Package for the Social Sciences (SPSS) version 1. Ferlay J, Soerjomataram I, Dikshit R, Eser S, Mathers 23 (SPSS UK Ltd, Woking, UK). The unpaired two-tailed C, Rebelo M, Parkin DM, Forman D, Bray F. Cancer Student's t-test was used to compare mean values between incidence and mortality worldwide: sources, methods and two groups. Data are presented as mean ± SD. P ≤ 0.05 major patterns in GLOBOCAN 2012. Int J Cancer. 2015; was considered statistically significant. 136:E359-86. 2. Kamisawa T, Wood LD, Itoi T, Takaori K. Pancreatic Abbreviations cancer. Lancet. 2016; 388:73-85. 3. Burris HA 3rd, Moore MJ, Andersen J, Green MR, ADCC, antibody-dependent cell-mediated Rothenberg ML, Modiano MR, Cripps MC, Portenoy RK, cytotoxicity; CDC, complement-dependent cytotoxicity; Storniolo AM, Tarassoff P, Nelson R, Dorr FA, Stephens CICs, cancer-initiating cells; CSCs, cancer stem- CD, Von Hoff DD. Improvements in survival and clinical like cells; GPI, glycosylphosphatidylinositol; IHC, benefit with gemcitabine as first-line therapy for patients immunohistochemistry; mAbs, monoclonal antibodies; with advanced pancreas cancer: a randomized trial. J Clin MFI, mean fluorescence intensity; SRB, Sulforhodamine Oncol. 1997; 15:2403-2413. B; TGF, transforming growth factor; TNBC, triple- negative breast cancer. 4. Moore MJ, Goldstein D, Hamm J, Figer A, Hecht JR, Gallinger S, Au HJ, Murawa P, Walde D, Wolff RA, Author contributions Campos D, Lim R, Ding K, et al; National Cancer Institute of Canada Clinical Trials Group. Erlotinib plus gemcitabine HM conceived and designed the experiments. GAP compared with gemcitabine alone in patients with advanced performed the experiments and data analysis. AGD and pancreatic cancer: a phase III trial of the National Cancer www.oncotarget.com 20005 Oncotarget Institute of Canada Clinical Trials Group. J Clin Oncol, 20. Sato T, Murakumo Y, Hagiwara S, Jijiwa M, Suzuki C, 2007; 25:1960-1966. Yatabe Y, Takahashi M. High-level expression of CD109 5. Von Hoff DD, Ervin T, Arena FP, Chiorean EG, Infante is frequently detected in lung squamous cell carcinomas. J, Moore M, Seay T, Tjulandin SA, Ma WW, Saleh MN, Pathol Int. 2007; 57:719-724. Harris M, Reni M, Dowden S, et al. Increased survival in 21. Bizet AA, Liu K, Tran-Khanh N, Saksena A, Vorstenbosch pancreatic cancer with nab-paclitaxel plus gemcitabine. J, Finnson KW, Buschmann MD, Philip A. The TGF- New Engl J Med. 2013; 369:1691-1703. beta co-receptor, CD109, promotes internalization and 6. Conroy T, Desseigne F, Ychou M, Bouche O, Guimbaud degradation of TGF-beta receptors. Biochim biophys Acta. R, Becouarn Y, Adenis A, Raoul JL, Gourgou-Bourgade S, 2011; 1813:742-753. de la Fouchardiere C, Bennouna J, Bachet JB, Khemissa- 22. Dong F, Liu F, Yan S, Liu X, Jiang Z, Liu J. Elevated Akouz F, et al; Groupe Tumeurs Digestives of Unicancer, expression of CD109 in esophageal squamous cell PRODIGE Intergroup. FOLFIRINOX versus gemcitabine carcinoma. Pathol Oncol Res. 2015; 21:1273-1275. for metastatic pancreatic cancer. New Engl J Med. 2011; 23. Emori M, Tsukahara T, Murata K, Sugita S, Sonoda T, Kaya 364:1817-1825. M, Soma T, Sasaki M, Nagoya S, Hasegawa T, Wada T, Sato 7. Rahib L, Smith BD, Aizenberg R, Rosenzweig AB, N, Yamashita T. Prognostic impact of CD109 expression in Fleshman JM, Matrisian LM. Projecting cancer incidence myxofibrosarcoma. J Surg Oncol. 2015; 111:975-979. and deaths to 2030: the unexpected burden of thyroid, liver, 24. Hashimoto M, Ichihara M, Watanabe T, Kawai K, and pancreas cancers in the United States. Cancer Res. Koshikawa K, Yuasa N, Takahashi T, Yatabe Y, Murakumo 2014; 74:2913-2921. Y, Zhang JM, Nimura Y, Takahashi M. Expression of 8. Ferlay J, Partensky C, Bray F. More deaths from pancreatic CD109 in human cancer. Oncogene. 2004; 23:3716-3720. cancer than breast cancer in the EU by 2017. Acta 25. Lin M, Sutherland DR, Horsfall W, Totty N, Yeo E, Nayar Oncologica. 2016; 55:1158-1160. R, Wu XF, Schuh AC. Cell surface antigen CD109 is a 9. Pillay V, Gan HK, Scott AM. Antibodies in oncolog y. N novel member of the alpha(2) macroglobulin/C3, C4, Biotechnol. 2011; 28:518-529. C5 family of thioester-containing proteins. Blood. 2002; 10. Scott AM, Allison JP, Wolchok JD. Monoclonal antibodies 99:1683-1691. in cancer therapy. Cancer immunity. 2012; 12:14. 26. Haun RS, Fan CY, Mackintosh SG, Zhao H, Tackett AJ. 11. Reichert JM., Dhimolea E. The future of antibodies as CD109 overexpression in pancreatic cancer identified by cancer drugs. Drug Discov Today. 2012; 17:954-963. cell-surface glycoprotein capture. J Proteomics Bioinform. 12. Modjtahedi H. Monoclonal Antibodies as Therapeutic 2014; 10:S10003. Agents: Advances and Challenges. Iranian journal of 27. Hagikura M, Murakumo Y, Hasegawa M, Jijiwa M, immunology. 2005; 2:3-20. Hagiwara S, Mii S, Hagikura S, Matsukawa Y, Yoshino 13. Pento JT. Monoclonal Antibodies for the Treatment of Y, Hattori R, Wakai K, Nakamura S, Gotoh M, Takahashi Cancer. Anticancer Res. 2017; 37:5935-5939. M. Correlation of pathological grade and tumor stage of urothelial carcinomas with CD109 expression. Pathol Int. 14. Shepard HM, Phillips GL, Thanos CD, Feldmann M. 2010; 60:735-743. Developments in therapy with monoclonal antibodies and related proteins. Clinical medicine. 2017; 17:220-232. 28. Zhang JM, Hashimoto M, Kawai K, Murakumo Y, Sato T, Ichihara M, Nakamura S, Takahashi M. CD109 expression 15. U.S. Food & Drug Administration. Hematology/Oncology in squamous cell carcinoma of the uterine cervix. Pathol Int. (Cancer) Approvals & Safety Notifications. Available 2005; 55:165-169. online: http://www.fda.gov/Drugs/InformationOnDrugs/ ApprovedDrugs/ucm279174.htm (accessed on 26th June, 29. Hagiwara S, Murakumo Y, Sato T, Shigetomi T, Mitsudo 2017). K, Tohnai I, Ueda M, Takahashi M. Up-regulation of CD109 expression is associated with carcinogenesis of the 16. European Medicines Agency. Available online: http://www. squamous epithelium of the oral cavity. Cancer Sci. 2008; ema.europa.eu/ema/index.jsp?curl=pages/special_topics/ 99:1916-1923. landing/cancer_disease_area.jsp&mid=WC0b01ac058034ed06 (accessed on 26th June, 2017). 30. Hasegawa M, Moritani S, Murakumo Y, Sato T, Hagiwara S, Suzuki C, Mii S, Jijiwa M, Enomoto A, Asai N, Ichihara 17. Modjtahedi H, Ali S, Essapen S. Therapeutic application of S, Takahashi M. CD109 expression in basal-like breast monoclonal antibodies in cancer: advances and challenges. carcinoma. Pathol Int. 2008; 58:288-294. Br Med Bull. 2012; 104:41-59. 31. Tao J, Li H, Li Q, Yang Y. CD109 is a potential target 18. Reichert JM. Antibodies to watch in 2017. MAbs. 2017; for triple-negative breast cancer. Tumour Biol. 2014; 9:167-181. 35:12083-12090. 19. Sutherland DR, Yeo E, Ryan A, Mills GB, Bailey D, Baker 32. Ohshima Y, Yajima I, Kumasaka MY, Yanagishita T, MA. Identification of a cell-surface antigen associated with Watanabe D, Takahashi M, Inoue Y, Ihn H, Matsumoto Y, activated T lymphoblasts and activated platelets. Blood. Kato M. CD109 expression levels in malignant melanoma. 1991; 77:84-93. J Dermatol Sci. 2010; 57:140-142. www.oncotarget.com 20006 Oncotarget 33. Dong F, Lu C, Chen X, Guo Y, Liu J. CD109 is a novel TGF-beta signalling and responses. Biochem J. 2016; marker for squamous cell/adenosquamous carcinomas of 473:537-547 the gallbladder. Diagn Pathol. 2015; 10:137. 43. Sakakura H, Murakumo Y, Mii S, Hagiwara S, Kato T, Asai 34. Zong G, Xu Z, Zhang S, Shen Y, Qiu H, Zhu G, He S, Tao M, Hoshino A, Yamamoto N, Sobue S, Ichihara M, Ueda T, Chen X. CD109 Mediates Cell Survival in Hepatocellular M, Takahashi M. Detection of a soluble form of CD109 in Carcinoma Cells. Dig Dis Sci. 2016; 61:2303-2314. serum of CD109 transgenic and tumor xenografted mice. 35. Jia W, Ren C, Wang L, Zhu B, Jia W, Gao M, Zeng F, PLoS One. 2014; 9:e83385. Zeng L, Xia X, Zhang X, Fu T, Li S, Du C, et al. CD109 44. Eckel-Passow JE, Lohse CM, Sheinin Y, Crispen PL, Krco is identified as a potential nasopharyngeal carcinoma CJ, Kwon ED. Tissue microarrays: one size does not fit all. biomarker using aptamer selected by cell-SELEX. Diagnostic pathology. 2010; 5:48-1596-5-48. Oncotarget. 2016; 7:55328-55342. https://dx.doi.org/10.18 45. Collisson EA, Olive KP. Pancreatic Cancer: Progress and 632%2Foncotarget.10530. Challenges in a Rapidly Moving Field. Cancer research. 36. Dong F, Wang J, Xu Y, Cheng Z, Chen X, Wang Y, Liu J. 2017; 77:1060-1062. CD109 expression is upregulated in penile squamous cell 46. Beers SA, Glennie MJ, White AL. Influence of carcinoma. Oncol Lett. 2017; 14:6012-6016. immunoglobulin isotype on therapeutic antibody function. 37. Ozbay PO, Ekinci T, Yigit S, Yavuzcan A, Uysal S, Soylu Blood. 2016; 127:1097-1101. F, Cakalagaoglu F. Investigation of prognostic significance 47. Ioannou N, Dalgleish AG, Seddon AM, Mackintosh D, of CD109 expression in women with vulvar squamous cell Guertler U, Solca F, Modjtahedi H. Anti-tumour activity carcinoma. Onco Targets Ther. 2013; 6:621-627. of afatinib, an irreversible ErbB family blocker, in human 38. Shiraki Y, Mii S, Enomoto A, Momota H, Han YP, Kato T, pancreatic tumour cells. Br J Cancer. 2011; 105:1554-1562. Ushida K, Kato A, Asai N, Murakumo Y, Aoki K, Suzuki 48. Modjtahedi H, Styles JM, Dean CJ. The human EGF H, Ohka F, et al. Significance of perivascular tumour cells receptor as a target for cancer therapy: six new rat mAbs defined by CD109 expression in progression of glioma. J against the receptor on the breast carcinoma MDA-MB 468. Pathol. 2017; 243:468-480. Br J Cancer. 1993; 67:247-253. 39. Emori M, Tsukahara T, Murase M, Kano M, Murata K, 49. Khelwatty SA, Essapen S, Seddon AM, Modjtahedi H. Takahashi A, Kubo T, Asanuma H, Yasuda K, Kochin V, Growth response of human colorectal tumour cell lines to Kaya M, Nagoya S, Nishio J, et al. High expression of treatment with afatinib (BIBW2992), an irreversible erbB CD109 antigen regulates the phenotype of cancer stem-like family blocker, and its association with expression of HER cells/cancer-initiating cells in the novel epithelioid sarcoma family members. Int J Oncol. 2011; 39:483-491. cell line ESX and is related to poor prognosis of soft tissue 50. Ioannou N, Seddon AM, Dalgleish A, Mackintosh D, Solca sarcoma. PloS One. 2013; 8:e84187. F, Modjtahedi H. Acquired resistance of pancreatic cancer 40. Yokoyama M, Ichinoe M, Okina S, Sakurai Y, Nakada N, cells to treatment with gemcitabine and HER-inhibitors is Yanagisawa N, Jiang SX, Numata Y, Umezawa A, Miyazaki accompanied by increased sensitivity to STAT3 inhibition. K, Higashihara M, Murakumo Y. CD109, a negative International journal of oncology. 2016; 48:908-918. regulator of TGF-beta signaling, is a putative risk marker 51. Khelwatty SA, Essapen S, Seddon AM, Fan Z, Modjtahedi in diffuse large B-cell lymphoma. Int J Hematol. 2017; H. Acquired resistance to anti-EGFR mAb ICR62 in cancer 105:614-622. cells is accompanied by an increased EGFR expression, 41. Chuang CH, Greenside PG, Rogers ZN, Brady JJ, Yang D, HER-2/HER-3 signalling and sensitivity to pan HER Ma RK, Caswell DR, Chiou SH, Winters AF, Gruner BM, blockers. Br J Cancer. 2015; 113:1010-1019. Ramaswami G, Spencley AL, Kopecky KE, et al. Molecular 52. Puvanenthiran S, Essapen S, Seddon AM, Modjtahedi H. definition of a metastatic lung cancer state reveals a Impact of the putative cancer stem cell markers and growth targetable CD109-Janus kinase-Stat axis. Nat Med. 2017; factor receptor expression on the sensitivity of ovarian 23:291-300. cancer cells to treatment with various forms of small 42. Li C, Hancock MA, Sehgal P, Zhou S, Reinhardt DP, molecule tyrosine kinase inhibitors and cytotoxic drugs. Int Philip A. Soluble CD109 binds TGF-beta and antagonizes J Oncol. 2016; 49:1825-1838.